Wander is the leading travel marketplace for exceptional vacation homes, combining hotel-grade amenities, inspiring views, pristine cleaning, and 24/7 concierge service – earning best-in-category guest satisfaction across thousands of locations from coast to coast.

As Wander continues expanding across thousands of destinations, reaching the right travelers at the right moment has become increasingly important. To accelerate growth and better connect with future guests, Wander partnered with Minerva to build a more intelligent and scalable performance marketing engine.

Over the past year, Wander and Minerva have collaborated to build a state-of-the-art performance marketing function leveraging Minerva’s platform. Throughout our work together, Wander has built a deeper understanding of their customers, improved targeting and creative personalization throughout the funnel and seen a ROAS lift of 3.4x.

Analysis

As a first step, Minerva integrated Wander’s first-party customer data with broader consumer intelligence to create a complete view of its audience. The analysis surfaced a key insight:

Wander’s highest-value guests were not one homogeneous audience, but a collection of distinct customer segments — from young urban professionals to affluent family travelers and other lifestyle-based cohorts. This created an opportunity to evolve targeting and creative strategies beyond broad audience approaches and better align messaging with the motivations of each traveler group.

Audience Optimization

Expanding from analysis to activation, Minerva developed a dynamic score for Wander's highest-value customers to power full-funnel targeting.

[IMAGE/TABLE]

The scoring model created a stronger foundation for optimization. Meta’s algorithms perform best when they receive high-quality conversion signals, and Wander had an opportunity to provide clearer indicators of who its highest-value travelers were. By improving audience definition and signal quality, campaigns could learn faster, identify more relevant users, and optimize toward outcomes that mattered most.

Minerva built the seed from the combination of Wander and Minerva’s data - every customer scored across 140+ consumer attributes (income, wealth, property ownership, life stage, life events, family composition). The top of that distribution became the dynamic high-value cohort.

Then we shipped the infrastructure to make that cohort actually move ROAS on Meta:

1. High-value lookalikes: audiences seeded from the top cohort, giving Meta scale against people it can recognize as similar to the actual Wander buyer profile
2. Key-attribute segments: direct targeting against people who match the cohort's defining attributes (HNWI signals, life-stage triggers), giving campaigns high-fidelity precision
3. Server-side CAPI + synthetic mid-funnel signals: clean attribution back to Meta that bridges the gap between top-of-funnel impressions and bottom-of-funnel bookings, so the algorithm can see what's converting and optimize against it

Together, these changes helped create more effective learning and optimization within Meta campaigns.

The results:

• ROAS improved from 1.69x to 5.18x within key audience segments
• High-value lookalike audiences achieved ROAS of 5.72x
• Campaign efficiency improved across the funnel

In-progress - Creative Optimization

The audience analysis surfaced more than a targeting gap. Once it mapped Wander's actual buyer cohorts — distinct clusters spanning young metro-area DINKs, affluent family travelers, and a wide tail of lifestyle sub-segments — the creative requirement followed.

Wander's existing slate was strong on brand and had carried them to scale, but it was built for a single brand-awareness moment served broadly. The cohort diversity the analysis revealed needed something different. The new creative had to:

• Match each distinct buyer cohort with persona-tuned variants, not one slate for everyone
• Span awareness, consideration, and conversion stages, not just top-of-funnel awareness
• Use messaging calibrated to each cohort's specific motivators, not generic luxury-travel language
• Refresh on a cadence that prevented audience fatigue on the existing slate

We ran our creative workflow to build it:

1. Brand context: research agents structured Wander's voice, positioning, brand pillars, and category into a canonical model — the source of truth every downstream agent consumed
2. Competitive intelligence: competitive-teardown agents indexed the luxury-travel ad landscape against the brand model, scored creative patterns by saturation and engagement, and surfaced the white-space wedge Wander could own
3. Format diagnostics: channel agents ran performance analysis across Meta + high-consideration travel: statics vs. video conversion rates, scroll-stop hook patterns by category, awareness-stage hook structures
4. Angle generation: angle agents produced a structured matrix of positioning angles across cohort × funnel-stage, each tuned to its cohort's actual motivators, running the brand model and competitive inputs in parallel
5. Copy generation: copy agents drafted awareness-stage ad copy from each angle, calibrated for the moment when the viewer doesn't yet know Wander
6. Creative production: production agents compiled copy and visual variants into a batch of static creatives, each indexed to a specific cohort × angle × stage combination